Over 20 years experience of training in the care sector

Call us on

020 3129 5667

How to use AI safely in adult social care

At your next team meeting, ask where staff already use AI in their work. Agree which tools and tasks are approved, what information staff can enter, and who checks AI-assisted work before anyone relies on it.

Three steps for managers

  • Set clear limits. List the tools and uses you’ve approved. Make clear that staff must keep identifiable care records, incident details and confidential staff information out of public or unapproved tools.
  • Assign a reviewer. Name someone who will check AI-assisted work against the original information, challenge errors or bias, and take responsibility for the final version.
  • Involve people early. If a tool will affect someone’s care, explain how it will be used in a way they can understand. Give them a chance to ask questions and express their preferences.

Follow the guidance for your service

For England, the Department of Health and Social Care covers responsible use and human oversight. Read Using AI in adult social care and the adult social care digital skills framework.

For Scotland, the Scottish Social Services Council sets out what responsible use means for workers and employers. Read the SSSC guide to responsible use of AI.

Involving people in decisions about their care

Explain what the tool does, where it might affect support and who makes the final decision. Ask what matters to the person, make it easy to raise concerns and follow the relevant consent and capacity requirements for your setting. For example, if an approved tool helps draft a care plan, check the draft with the person wherever possible so it reflects their preferences and routines.

Handling personal information lawfully

Treat the conversation about using AI in someone’s care and the decision about how to process their information as separate steps. For each proposed use of personal information, establish and document a UK GDPR lawful basis. Health information also needs a special-category condition. Check what the supplier does with information, how long it keeps it and who can access it. Carry out a data protection impact assessment before processing that is likely to result in a high risk to people.

Removing a name still leaves a risk of identification if other details point to the person. Give staff a straightforward route to ask before using real care information. Read the ICO guidance on lawfulness in AI and when to carry out a data protection impact assessment.

Give staff examples they can use on shift

  • A generic team briefing. An approved tool can help draft a briefing from non-confidential material. Check the wording and facts before sharing it.
  • An incident or care plan. Keep identifiable information in the authorised environment. If your organisation has approved AI for this specific task, compare the output against the original record, check for missing context and ensure a named member of staff signs it off.

AI for workforce learning in CareTutor Pathway

Pathway’s AI Course Builder is available for organisations creating their own eLearning. It can help draft learning content from a source document or from scratch. Have a subject lead check the draft for accuracy and relevance before assigning it to staff. Explore CareTutor Pathway LMS.

Facebook
Twitter
LinkedIn